Stable Diffusion

Good settings for blending things together in img2img

Mask blur: 4
Masked content: Original ! Use a crude self-drawn base to seed the diffusion.
Type: Euler a
Sampling steps: 50
Denoising strength: 0.35 - 0.65 ! Play with this variable, start at the extremes.

Windows 10 Powershell launch with network availability

PS D:\_STABLEDIFFUSION\StableDiffusion\stable-diffusion-webui-XL> ./webui.bat --listen

Gotchas

Error: "modules.devices.NansException: A tensor with all NaNs was produced in Unet. This could be either because there's not enough precision to represent the picture, or because your video card does not support half type. Try setting the "Upcast cross attention layer to float32" option in Settings > Stable Diffusion or using the --no-half commandline argument to fix this. Use --disable-nan-check commandline argument to disable this check"
Do a txt2img inference first before img2img, or you get this error.

Prompts

Painting Wallpaper upscale

A vibrant, ultra-detailed photo-realistic painting of a landscape scene, bursting with rich, vivid colors and dramatic lighting. The composition features intricate textures, strong contrast, emphasizing beauty and visual appeal. The scene is harmonious yet striking, captivating the viewer with its bold yet refined details. Inspiration taken from John Howe, Alan Lee, and Ilya Nazarov.

Settings

Resize to scale: 1
Sampling: 60, DPM++ 2M Karras
CFG Scale: 12
Denoising: 0.1
SD upscale script
Scale factor: 2
Upscaler: ESRGAN_4x